No abstract

Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-pro t purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. Impact ionization rates for electrons and holes in three semiconductors with particular band structure characteristics are examined to determine underlying factors influencing their qualitative behavior. The applicability of the constant matrix element approximation is investigated, and found to be good for the indirect gap material studied, but overestimates threshold softness in the direct gap materials. The effect that final states in the ⌫ valley have in influencing characteristics of the rate in the direct gap materials is investigated, and it is found that they play a significantly greater role than the low density of ⌫ valley states would suggest. The role of threshold anisotropy in affecting threshold softness is examined, and it is concluded that it plays only a small part, and that softness is controlled mainly by the slow increase in available phase space as the threshold energy is exceeded.

Thin film multilayered spin glass CuMn/Cu structures display glassy dynamics. The freezing temperature, T f , was measured for forty layers of CuMn films of thickness L = 4.5, 9.0, and 20.0 nm, sandwiched between non-magnetic Cu layers of thickness ≈ 60 nm. The Kenning effect, T f ∝ ℓn L, is shown to follow from power law dynamics where the correlation length grows from nucleation as ξ(t, T ) = c1a0(t/τ0) c 2 (T /Tg ) , leading to [(T f /Tg)c2ℓn(tco/τ0)] + ℓnc1 = ℓn(L/a0). Here, Tg is the bulk spin glass temperature, c1 and c2 are constants determined from the spin glass dynamics, tco is the time for the correlation length to grow to the film thickness, τ0 is a characteristic exchange time ≈ /kBTg, and a0 is the average M n − M n separation. For t ≥ tco, the magnetization dynamics are simple activated, with a single activation energy ∆max(L)/kBTg = (1/c2)[ℓn(L/a0) − ℓnc1] that does not change with time. Values for all these parameters are found for the three values of L explored in these measurements. We find experimentally ∆max(L)/kB = 907 K, 1,246 K, and 1,650 K, respectively, for the three CuMn thin film multilayer thicknesses, to be consistent with power law dynamics. We perform a similar analysis based on the activated dynamics of the droplet model, and find a much larger spread for ∆max(L) than found experimentally.

Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-pro t purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. An algorithm for calculating impact ionization rates in the semiclassical Fermi's Golden Rule approximation which is efficient close to threshold is presented. Electron and hole initiated rates are calculated for three semiconductors with particular band structure characteristics, as are the distributions of the generated carriers. Simple analytic expressions of the form RϭA(EϪE 0 ) P are fitted to the calculated rates. The role of the matrix elements in influencing the distribution of final states is investigated. In the direct gap materials, they act to significantly enhance the low-q transitions, while in the indirect gap case they have a lesser effect on the distribution. Results for GaAs obtained here and by several other workers are compared and possible causes of the discrepancies examined, including differences in band structure and approximations made in evaluation of the matrix element. It is found that these differences do not influence the rate sufficiently to account for the wider variation between authors, and so it is concluded that differences in the implementation of the rate integration algorithm are the main cause.